Machine Vision-Based Detection of Surface Defects of 3D-Printed Objects

Due to advances in 3D printing technologies, 3D object manufacturing has attracted significant attention nowadays. The market size of 3D printing is increasing exponentially, ranging from tiny toys to nuclear reactors. The significant advantage of this 3D print manufacturing over conventional manufacturing tools is that it can produce a complex object within a short period with a flexible but precise design. On the other hand, this tool has shown disadvantages as well. One of the main disadvantages is forming irregularities and defects within the 3D objects, which cost a significant amount of time and resources. Now the main challenge is to detect the defects on time and find a suitable solution, which might save a lot of time and money. In this study, a machine learning (ML) technique with a 3D vision camera is employed to detect and classify the defects of 3D objects. First, images collected from a depth (3D) camera are utilized for training the model. Then, the trained model is tested on a large set of real objects to detect defects. With the application of this technique, it is possible to detect defects while printing. As surface defects are a serious issue for 3D printed products and many other types of manufacturing methodologies, we hope the research outcome can be applied in different manufacturing areas to maintain the pavement to the advanced inspection on time with high-quality accuracy.